-19-

PASSAGEAND LOCAL • OF REDSHANKSTringa totanus ON rl•{ECLYDE ESTUARY AS SHOWNBY DYE-MARKING by Robert W. Furness and Hector Galbraith

The inner Clyde estuary is an internationally important winter feeding ground for ducks and waders. It ranks third in for peak winter wader numbers(Standring 1978). Its n•Dstimportant asset is a flock of RedshanksTringa totanus of up to 10,000 birds, or roughly 8% of the west European population in n•Dst recent winters (Gibson 1978), •it-• smaller flocks of OystercatchersHaemato•us ostralequs, Lapwings Vanellus vanellus, DunlinsCalidris alpina, Curlews Numeniusarquata and GreenshanksTrin•a nebularia. The waders achieve a density per square kilometre of mudflat which is exceeded in Scotland only on the Ythan estuary (calculated from Birds of Estuaries Enquiry data). Much of the inner estuary has already been reclaimed• Nearly 50% of the high water line is delimited by retaining sea walls with n•Dtorways, housing or industrial developments in•nediately behind. Further reclamation is in progress although at present on a small scale. As in the Forth (Furness 1973 a, b; Pienkowski & Clark 1979) waders on the inner Clyde estuary have a limited number of roost sites, mainly on 'unnatural' constructions, and are subject to considerable disturbance from aircraft, ships, trains, lorries and people (Goudie 1980).

Numbers of Redshanks on the estuary in winter 1978 - 79 were unusually low, even before the onset of cold weather. Limited sampling of intertidal invertebrate densities (Broadway 1979) suggested that while numbers of Nereis diversicolor were the same as in previous winters, densities of Coro•)hiumvolutator were an order of magnitude below normal. Co.ro•hiumforms the bulk of Redshanks'diet on the Clyde, and is the preferred food in manyother estuaries (Goss-Custard 1977). It is also noteworthy that Dunlin numbers on the estuary were only 20% of normal in 1978-79 - they are thought to feed largely on Coro•hiumon the Clyde (I.Gibson pers. cc•m. ). The disappearance of about 5000 Redshanks gave cause for concern as it represents 4% of the west European population. The situation was particularly interesting because of the previous high fSdelity of local ringed Redshanks to the Clyde (Mackie 1976). The 21 recoveries in winters up to 1978-79 of Redshanks ringed in winter on the Clyde estuary had all been on the Clyde Estuary. In winter 1978-79 there were four recoveries of birds ringed in previous winters; two on the Clyde, one in Holy Loch, Argyll and the other at Culmore, Londonderry,Northern Ireland giving a clue to the newwintering area of birds deserting the Clyde. For obviousreasons Redshanks became the target speciesfor cannon-nettingin 1979-80.The questions we wished to examine included: is the Clyde population hc•ogenous, with all the birds using all parts of the estuary? Is there any turnover through the area in winter (eg• Evans 1979)? Howmuch passage occurs in autumnand spring? These questions are very similar to those posed by Pienkowski and Clark (1979) and as in their study we aimed to use dye-marking to provide answers.

Results and discussion

Throughout most of the winter nearly all the Clyde estuary Redshanksroost in a single-species flock on a motorway wall at Langbank (Fig. 1). From there they can be movedto West Ferry shingle spit without difficulty. A catch of 1057 unringed birds, 64 retraps and 3 controls was taken on 20 October 1979. It was already clear by this date that Redshank numbers were below the normal in 1979-80. Only half the expected number were present. Three features of the catch were of particular interest: it contained very few first year birds (14 out of 521 aged were 1st year; i.e. 2.7%). Subsequentonbservations at roosts confirmed this was representative of the whole flock. Smaller catches in 1979-80 at Ardrossan, Ayrshire and on the Ythan and Forth estuaries (S.R.Baillie, M.W•Pienkowskipers. c•.) contained much larger proportions of young birdsø Mist net catches on the Clyde estuary between 1973 and 1976 contained 15%juveniles in August, 18% in September and 49% in October (samples of 76,22 and 71; Mackie 1976). While mist net catches may be biased in favour of juveniles (Pienkowski and Dick 1976),they are comparable to the proportions found in cannon net catches in other estuaries. Swann (1979) concluded from samples cannon-netted in the Moray Firth basin that the median Redshankcompleted mDult by 10 October. Mackie (1976) found an almost identical pattern in the Clyde. Moult scores were recorded from a randomsample of 136 birds in our catch. Less than half had completed moult on 20 October (Fig.2) suggesting that our birds were at least 10 days later than normal in their moult. Weights of 184 birds chosen at random averaged 149.8g (s.d. 11.9, s.e. 0.88). Contrary to the pattern described by Davidson (1979a) for other areas live weights of Redshankson the Clyde estuary (Mackie 1976) show little clear change between September and February, with a suggestion of a winter peak in Januaryø The mean of a sample of 215 birds from a cannon net catch on 11 November 1973 was 155.0g (s.d. 11o8, s.e. 0.80). The sample caught on 20 October 1979 was significantly lighter by over 5g. Measurementsof the protein reserve of 112 individuals using the method described by Davidson (1979b) showed that all the birds were in poor condition, most having a muscle condition index in the upper range of those which were found starved to death on the Ythan in January-February 1979 (Baillie 1980), and well below that of 'normal' birds (Davidson pers.cc•m.)o Lipid levels were also estimated to be below normal (ca 6•7% comparedto ca 8% Davidson pers. cc•m. and 1979a). The poor body condition, late moult, low pro.portion of juveniles and low numbersall suggest that conditions on the Clyde estuary are not now as suitable for wintering Redshanksas they once were.

Before releasing the birds we dyed 333 individuals with picric acid on the breast and underparts. Counts at Redshank roosts of the numberspresent and the proportion of dye-marked individuals in sampleswere madeas often as possible. We also visited most of the Redshank roosts between Helensburgh (north) and Troon (south) to look for marked Redshanksleaving the Clyde estuary. No counts were madeat low tide but dye-markedbirds were seen feeding over all the mudflats of the inner estuary.

Redshankroosts in the Clyde estuary are shownin Figure 1. In September and in October before the cannon net catch, 4000-5000 Redshanks roosted on the wall at Langbank• Flocks of less than 100-300 birds were found in Holy Loch, Hunterston Bay and Bogside Flats. Counts of Redshanksat the main inner Clyde roosts ( to Dumbartonand to Langbankbut not extending to Woodhall or Ardmore) after the catch are given in Table 1, together with numbersof dye-marked birds in sample counts. Proportions of dye-marked birds did not differ significantly between these roosts.

In•nediately after the catch a small numbermoved to roost to Ardmore or Woodhall. Half the flock moved to East Ferry or the Esso wall. Counts of dye-markedbirds at Woodhall and Ardmore (Table 2) showthat there are 'distinct local Redshankpopulations at these two sites. Small numbersjoined these because of the disturbance of cannon-netting but the proportion of markedbirds fell back to zero by December.These birds clearly returned to Langbankleaving only the local populations at Woodhall and Ardmore• No dye-marked Redshankswere found at any of the other roosts between Helensburgh and Troon (Table 3)• The seasonal pattern in the inner Clyde estuary showeda constant proportion of dye-marked birds at these roosts until 5 March 1980, Clearly there was no significant movementof Redshanksto the Clyde between 20 October 1979 and 5 March 1980 otherwise the proportion dye-markedwould have decreased. This agrees with the high site-fidelity in winter indicated by ringing data. As the proportion dye-marked remained constant we can estimate the population size at capture by simple Lincoln index. Between 22 October 1979 and 5 March 1980, 338 observations of dye-marked birds were made in samples totalling 4482. As 333 birds were dye-marked we estimate a flock -20-

Table 1. Counts of Redshanks and proportions of dye-marked birds in samples observed at the main roosts on the inner Clyde estuary.

Date Observer Roost Count Dye-markedin sample total flock

20-10-79 RWF Langbank 4500 4500 22-10-79 RWF Langbank 2500 57/728=7.8% East Ferry 2200 9/145=6.2% 4700 23-10-79 RWF Langbank 2500 15/214=7.0% East Ferry 1500 13/200=6.5% Esso Wall 520 4520 5-11-79 RWF Langbank 3800 East Ferry 650 5/80 =6.3% 4450 15-11-79 RWF Langbank 3000 East Ferry 500 29/423=6.9% 3500 20-11-79 HG Langbank 3000 21/268=7.8% 3000 29-11-79 HG Langbank 17/249=6.8% 3-12-79 HG Langbank 16/203=7.9% 10-12-79 HG Langbank 16/200=8.0% 30-12-79 HG Langbank 14/197=7.1% 5-1 -80 HG Langbank 3000 15/148=10.1% 3000 16-! -80 HG Langbank 11/113=9.7% East Ferry 0/31 =0% 3-2 -80 }• Langbank 9/87 =10.3% 15-2 -80 RWF Langbank 3100 52/636=8.2% 3100 5-3 -80 RWF Langbank 2200 26/320=8.1% Milton 750 2950 15-3 -80 RWF Milton 4200 42/800=5.2% Longhaugh 1100 5300 16-3 -80 HG Langbank 9/200=4.5% 18-3 -80 HG Milton 3500 Langbank 2500 9/160=5.6% 6000 23-3 -80 RWF Langbank 3500 Milton 2500 19/400=4.8% 6000 28-3 -80 RWF Milton 4000 9/300=3.0% Langbank 2600 East Ferry 50 6650 30-3 -80 RWF Esso Wall 2300 Milton 650 Langbank 150 5/100=5.0% 3100 5-4 -80 RWF Milton 900 Esso Wall 700 Langbank 250 East Ferry 100 1950 13-4 -80 RWF Longhaugh 100 Langbank 32 East Ferry 11 143 4-5 -80 RWF Langbank 7 Milton 10 17

,Ikm,

Figure 1. The main Redshankroosts on the inner Clyde estuary which were used in winter 1979-80: A, Ardmore; B, •ilton; C, Esso Wall; D, Longhaugh; E, East Ferry; F, West Ferry; G, Langbank Wall; H, Woodhall. -21-

Table 2. Counts of Redshanksand proportions dye-marked at Woodhall and Ardmore on the lower part of the Clyde estuary.

Roost Date Observer Count Dye-marked

Ardmore 26-10-79 RWF 140 3/140=2.1% 20-12-79 HG 34 0/34 =0.0% 17-2-80 RWF 38 0/38 =0.0% Woodhall 23-10-79 RWF 171 8/171=4.7% 27-10-79 RWF 24 0/24 =0.0% 5-11-79 RWF 12 0/12 =0.0% 20-11-79 HG 60 1/60 =1.7% 29-11-79 HG 91 4/91 =4.4% 9-12-79 HG 58 0/58 =0.0% 15-2-80 RWF 35 0/35 =0.0%

Table 3. Counts of Redshanks and numbers dye-marked at roosts away from the inner Clyde estuary.

Date Observer Roost Count Dye-marked

27-10-79 RWF Gourock 16 0 Inverkip 26 0 16-1-80 RWF Gourock 24 0 Inverkip 37 0 Ardrossan 29 0 Hunterston 155 0 16-2-80 HG/RWF Irvine 60 0 Ardrossan 20 0 Hunterston 220 0

N

2O

0 Figure 2. Distribution of moult scores in a sample of 136 Redshanks caught at West Ferry on 20 OctobeI 35 4O 45 50 1979. The median moult score is 49. Moult score size of 4350 birds. This gives us some confidence as our field counts were of 4500-4700 around the date of the catch.

Numbersof Redshankson the inner Clyde roosts fell during the winter. This cannot be accounted for by inland feeding at high tide as Redshanks on the Clyde do not move to the fields. Clearly some birds left the area between early Novemberand mid-February. We have little idea where these went. They do not seem to have joined the small Redshank populations of the nearby coast. Three sightings of dye-marked, not necessarily different, individuals were reported from the Firth of Forth where there were intensive observations of waders, including Redshanks, dyed with a different colour. Many could have movedto the Solway or Ireland where observer cover is low, but no dye-marked individuals were reported from other estuaries (but none were reported on the Clyde except by Clyde Ringing Group members!).

The dye began to be lost in March as birds moulted body feathers, but we were able to approach close to roosting flocks to count dye-marked individuals and we were certain that fading was not severe enoughto be a problem until April when counting was abandoned. The arrival of spring passage Redshanks was coupled with a decrease in the percentage dye-marked from 7.7% (winter average) to 3.5% (average for 28 and 30 March). The proportion dye-marked is close to that predicted by an arrival of 3600 new birds with no loss of the winter residents (3.4%) suggesting that this is what happened. Most birds left between 28 March and 13 April, so the spring passage birds were probably on the estuary for no more than 15 days.

We feel that the movementof Redshanks out of the estuary in mid-winter is unusual and may reflect poor local conditions in winters 1978-79 and 1979-80. Densities of Corophiumappear to have been similar in both winters (Goudie 1980) but reasons for these abnormally low levels are obscure. Wewait to see whether Corophiumrecovers by winter 1980-81 and how the Redshankswill respondøOur aim for 1980-81 will be to obtain more information on the ages and racial origin of Redshankspassing through the Clyde in autumnand spring, and to look at the survival body condition and feeding behaviour of our winter flock.

Acknowledgements

For access to private land we thank Mr Guy McKinlay. The catch could not have been madewithout Les Goodyerto whomwe are very grateful. We thank Keith Brockie and Mike Nicoll (Tay R.G.), Fraser Symonds(N.C.C. Edinburgh), Mike Moser and Shirley Jones (DurhamUniversity), Bridget Furhess (AberdeenUniversity), TomByars and John McKee(Clyde R.G.) for help in cannonnetting, Dr P. Monaghanand the Tay R.G. for lending us cannonnets, and Dr P.R.Evans for advice and encouragement.Nick Davidsonanalysed measurementsof musclereserve We thank him, Iain Gibsonand Dr Peter Mackie for providing unpublisheddata and for constructive criticisms. GlasgowUniversity provided logistic and financial assistance for this work. Finally we thank the Scottish Ornithologists Club for a grant towards the purchase of cannon-nets to allow us to continue this study. -22-

References

Baillie,S. 1980. The effect of the hard winter of 1978/79 on the wader populations of the Ythan estuary. Wader Study Group Bulletin 28:15-17 Broadway,K. 1979. The effect of avian predation on the benthic macroinvertebrates of the Clyde estuary. Unpubl. B.S.c thesis, University of . Davidson,N.C. 1979a. Decline in the condition of Redshanksduring the winter. Wader Study Group Bulletin 26:15 Davidson,N.C. 1979b. A technique for protein reserve estimation in live RedshanksTringa totanus. Wader Study Group Bulletin 27:14-15 Evans,P.Ro 1979. Shorebird research on the Tees estuary,N.E.England. Wader Study GroupBulletin 27:21-22 Furness,R.W. 1973ao Wader populations at Musselburgh. Scott. Birds 7:275-281 Furness,R.W. 1973b. Roost selection by waders. Scott. Birds 7:281-287 Gibson,I• 1978. Numbers of birds on the Clyde estuary. In Nature Conservation interests on the Clyde estuary. NCC:Balloch. Goss-Custard,J.D. 1977. The energetics of prey selection by Redshank, Tringa totanus (L), in relation to prey density. J.Anim. Ecol. 46:1-19 Goudie,K.J•N. 1980. Feeding by Redshank, Tringa totanus and Lapwing, Vanellus vanellus on the inner Clyde estuary. Unpubl. B.S.c. thesis, University of Glasgow. Mackie,P. 1976. A short note on Redshankin the upper Clyde estuary. Wader Study Group Bulletin 17:5-10 Pienkowski,M.W. and Clark,H. 1979. Preliminary results of winter dye-marking on the Firth of Forth, Scotland. Wader Study Group Bulletin 27:16-18 Pienkowski,M.W. and Dick,W.J.A. 1976. Sc•nebiases in cannon and mist netted samples of waders. Ringing and Migration 1:105-107 Standring,K. 1978. Winter birds of the Clyde estuary: a comparison with other estuaries. In Nature Conservation interests in the Clyde estuary NCC:Balloch Swann,R.L. 1979. Primary moult of Redshanks in the Moray Firth. Highland Ringing Group Report 3:16-20

R.W.Furness, Zoology Department, Glasgow University, Glasgow G12 8QQ, Scotland. H.Galbraith, 96 Neilston Road, Paisley, PA2 6EL, Scotland.

NON-RANDOMDISTRIBUTION IN ROOSTINGFLOCKS OF WADERSMARKED IN A CANNONNET CATCH. by Robert W. Furness and Hector Galbraith

In autumn 1979 some 4500-4700 Redshanks Tringa totanus roosted in a single compact flock on the inner Clyde estuary, Scotland. On 20 October 1979 we caught 1124 (24%) of these in a cannon net. The catch represented a complete section of the roosting flock from high on the shore to the water's edge. Almost all were adults (507 of 521 which were aged = 97%). During ringing 333 birds were chosen at random and dye-marked'with picric acid. Subsequent observations showed that most of the flock remained on the estuary until late March and no other Redshanksmoved into the area until spring passage in mid-March (Furhess and Galbraith 1980).

The flock used a number of roosts in the area, probably depending on the extent and nature of disturbance factors in this highly developed estuary. Usually more than 95% of the flock was to be found on only one roost. On three occasions (22 October, 15 November 1979 and 15 February 1980) ideal weather conditions, combined with luck, enabled very close observation of a large flock at the main roost.

Observations were made using 10x40 Leitz binoculars from a vehicle parked behind amotorway crash barrier less than 10m from the roosting birds at Langbank, Renfrewshire. As a result it was possible to count samples of the flock and record the number of dyed individuals in each group. Counts were made starting at one end of the flock. The first ten birds were c•unted and the number of dye marked individuals recorded. The next ten were counted and so on. On each occasion the count terminated long before reaching the end of the flock because more birds arrived or the flock was disturbed and some took off..

Results

A null hypothesis that the proportion of dyed birds in the three counts was the same gave expected counts of 56, 33 and 49 dyed individuals. Actual counts were 57, 29 and 52 respectively. The results do not differ significantly (x2 = 0.7) so can be combined to give a single sample. Overall, 138 of 1787 observations were of dye marked birds so the probability of any one bird being dye-mark½d was 0.0772. Groups of ten birds should contain an average of 0.772 dyed individuals. The expected numberof groups with 0, 1, 2 ...... 10 dyed individuals can be calculated from the Poisson distribution. If the dyed birds are randomlydistributed our observedcounts shouldmatch this series. Actual counts and the e•ected number (under the hypothesis of random distribution) are shown in Table 1. The hypothesis is clearly rejected (X•-= 15.7, p 0.01), showing that dyed birds were not randomly distributed through the roosting flocks. Instead they tended to occur in clumps more often than expected by chance.

Table Observed and expected frequencies of dye-marked birds per group of ten in }•dshank roosts at Langbank.

Numberdye-marked in Expected frequency Observed f•equency x2 group of ten if randomly distributed

0 81.8 97 2.8 1 63.1 40 8.5 2 24.4 28 0.5 3 6.3 9 1.2 4 or more 1.2 3 2.7 '•'•7 •- 177 15.7

Discussion

It is difficult to understand what these results mean. The discrepancy arises because there were more groups with no dye-marked birds or with several dye-marked birds and fewer than expected with only one. Pienkowski and Dick {1976) pointed out that cannon net catches may be biased because roosting waders often segregate into age and moult groups. In this study segregation must be within-flock as only one flock is involved. The birds continued to be non-randomly distributed within the roosting flock even four months after dye-marking so the behaviour is clearly stable over long periods of tire.